How it's made

AI Motion Design for YouTube: Production Breakdown

The Challenge: 30 Minutes of Motion Design on a YouTube Budget. Traditional motion design workflows, especially those built around After Effects, work great for shorter projects with healthy budgets. But a skilled motion designer can easily eat up the entire budget of a YouTube episode before you even get to the edit.Then our director suggested a different approach: what if we replaced most of the traditional After Effects workflow with ChatGPT? For generating custom visual assets, AI had become a much faster way to get exactly what we needed than drawing everything from scratch. And with 30 minutes of content to fill, that difference was huge.

Client

GFM News

Location

Los Angeles, DTLA

category

Youtube Content

director/dp

NOVSKII

Camera

Blackmagic 4k

‍The Reference: Raw YouTube News, Not Polished Television

Our client shared a reference inspired by the Morning Brew style of news content: presenters filming out on the streets of New York, with real city life happening behind them and motion graphics popping up throughout the episode.

That's one thing we love about YouTube. Content doesn't have to look like a polished TV broadcast to work. Sometimes the raw, unpredictable, slightly messy stuff is exactly what makes it engaging.

Luckily, our client understood that too.

We could've followed the reference pretty closely, and everyone would've been happy.

But we wanted to take it further.

Instead of dropping random graphics over the footage, we wanted to make the motion design feel like it actually belonged in the scene. Every visual needed to match the client's style, communicate the context, and ideally give the presenter something to interact with.

Designing the Motion Before Shooting

We started building the motion design into the script itself.

For example, we introduced balloons as physical storytelling elements. The presenter could inflate a balloon at the beginning of a topic, then release it, pop it, or puncture it at the end, depending on what the story was about.

That gave us something real to work with in post-production.

We also planned shots specifically for AI-generated visual integrations.

Using wide reference images of our Los Angeles rooftop location, we created expanded versions of the environment that allowed us to pull off dramatic zoom-outs, reveal the cityscape, and introduce oversized objects directly into the scene.

Instead of treating motion graphics as a separate layer, we were designing the footage and the future graphics together.

Building a New AI Motion Design Workflow

Once filming was done, we moved into Photoshop, using its AI-powered expansion tools to prepare backgrounds and extended environments for the upcoming generations.

Then came ChatGPT.

We used it to generate custom objects that matched the lighting, camera angle, perspective, and composition of our actual footage.

This was one of the biggest advantages over traditional stock assets.

Rather than spending hours searching for something that almost worked, we could describe exactly what we wanted and generate an object that already belonged in the shot.

With the right prompts, AI could pick up the scene's lighting and produce surprisingly consistent results.

Then we'd simply ask it to remove the background.

And just like that, we had a custom 2D asset ready to drop onto the timeline.

No manual illustration. No endless stock searches. No rebuilding every little object in After Effects.

Of course, generating an image was only part of the process. We still needed to make it work within the edit.

Making Static Graphics Feel Alive

The next step was animation, compositing, and timing.

We added movement to the generated assets, adjusted their placement, created masks where objects interacted with the presenter, and used fast zooms and perspective changes to keep the visuals moving.

Our goal was simple: make a 30-minute financial news episode engaging enough for an audience used to fast-moving, short-form content.

That meant constantly finding new ways to illustrate the story without turning the whole video into visual chaos.

And sometimes our editor got bored with static graphics.

That's when Kling AI came in.

We started feeding static plates into Kling to generate things like Bitcoin-shaped buildings growing out of the Los Angeles skyline and all kinds of other ridiculous visual moments.

Would a traditional motion artist build something like that for a YouTube news episode?

Probably not. And for good reason.

The time and cost would make absolutely no sense.

But with Kling, we could often get a usable animated plate in roughly ten minutes, including a few revisions.

Suddenly, ideas that would've been too expensive or time-consuming to even consider became realistic options.

The Hardest Part Wasn't Generating. It Was Choosing.

Here's something we didn't expect.

We went through several editors during the project.

Some ran into software and hardware limitations. Others had no trouble generating AI visuals, but the results just weren't right.

And that revealed something important about AI production.

Everyone can generate. Not everyone has the same taste.

Our director, Phil, rejected plenty of AI-generated assets. Not because they weren't realistic enough, but because they didn't fit the visual language of the project.

We weren't chasing photorealism.

In fact, some of the slightly cartoonish qualities of AI-generated objects worked better for the style we were developing.

What mattered was consistency, composition, timing, and whether the visual actually helped tell the story.

That's where creative direction becomes even more important.

When you can generate almost anything in seconds, the real skill is knowing what deserves to make it into the final cut.

And for simple 2D illustrations and clip-art-style graphics, AI was especially hard to beat. Assets that might take a motion artist a significant amount of time to prepare could now be generated in seconds.

Two Weeks Later: A Finished 30-Minute Episode

We completed the full 30-minute YouTube news episode in two weeks of editing and post-production.

And that included the time spent developing, testing, and refining the workflow itself.

We combined real cinematography, AI-generated motion graphics, custom visual assets, animated environments, compositing, and creative editing into one consistent visual style.

More importantly, we came out of the project with something we can use again.

A repeatable AI motion design workflow built for real YouTube production budgets.

Not just a collection of prompts or random AI tricks, but a production process that lets us create ambitious visuals without building a massive post-production team.

And it still has to meet our standards.

Because making something faster doesn't mean much if it doesn't look good.

‍
Ai motion design

‍
So, Is AI Killing Motion Designers? For certain types of production, it's definitely changing the economics. Especially for YouTube creators, news channels, educational shows, and other long-form projects that need a lot of visual storytelling without a Hollywood-sized budget.But AI doesn't automatically make a video good.It doesn't decide which visual joke lands, where the camera should move, when to cut, or whether a ridiculous Bitcoin building actually makes the story better.That's still the creative part. And that's exactly where we believe the real value of modern video production is heading. Less time making assets. More time making things worth watching.

Push to play

Let's work together!

Call now and get
a short script – for free!

Call now
CTA Concept Image

From strategy and storytelling to cinematic production and post. Just Click-N-Run.

runngun.media@gmail.com+1 (310) 925 3663
Social Icon

Get spammed

Back To Top Arrow